Download Versatile Cooperative Ligand Effects in Group 9 Transition Metal

Versatile Cooperative Ligand Effects in Group 9 Transition Metal Catalysis.
Applications in Transfer Hydrogenation & Hydrogen Autotransfer Reactions,
Ketene & Ketene Imine Synthesis and Hydroformylation
Z. Tang
Summary
Summary
Chemical production helps to sustain the modern society by providing people with energy, medicine
and materials. Catalysis plays a central role in chemical processes, as it serves to provide costefficient, “green” and selective pathways for otherwise highly energy-consuming, environmentallyhazardous or even inaccessible routes. Hence, catalyst development is one of the main aims of
modern chemical research. In the branch of homogeneous catalysis, the development of new catalytic
systems strongly relies on the design and synthesis of new organic ligands used to steer the reactivity
of transition metal complexes. In conventional organometallic catalysts, ligands usually act as
spectators during the reaction, exerting effects mainly by their electronic and steric properties, with
the metal center solely interacting with the substrates and mediating the reactions. Bifunctional
mechanisms in enzymes, catalysts in nature, that rely on more cooperative and active participation of
the organic framework around the metal ion, have inspired chemists to develop new types of ligands
that can facilitate chemical processes by directly participating in substrate activation processes, in
synergy with the function of the (transition) metal. The cooperative effects of these ligands provide
new patterns for catalysis that help to establish more cost-effective and efficient pathways for the
synthesis of chemical products. While cooperative ligands functioning as internal (Brønsted)
bases/acids in the reactions have been extensively applied and studied and ligands showing radicaltype reactivity are currently actively explored in catalysis research by several research groups, ligands
that are able to directly assist in hydride (nucleophile group) transfer in substrate activation have
remained quite underdeveloped, and only recently attracted attention of chemists. Representative
examples of the three types of cooperative ligands in (dehydrogenation)-related catalytic reactions are
described in Chapter 1. As the development of cooperative ligands capable of facilitating hydride
transfer processes could potentially have a profound impact on catalysis, the initial aim of the
research described in this thesis was to develop rhodium and iridium complexes containing such
cooperative ligand systems and to study their application in (de)hydrogenation-related catalysis.
Secondary picolylamine-type ligands coordinated to rhodium and iridium have been described to
convert rather easily to the corresponding “electrophilic” imine ligand initiated by NH deprotonation.
This indicates a quite strong reducing capability of these ligands, presumably associated with their
“hydride donating” ability. This was the starting point of the studies described in this Thesis.
In Chapter 2 the preparation of iridium and rhodium complexes 1 and 2 (Figure 1) based on the Nmethyl-1-(pyridin-2-yl)methanamine (Me-pma) ligand and their chemistry upon NH deprotonation of
the ligand are presented. While the iridium amido complex 1a could be obtained upon treatment with
one equivalent of base, the rhodium complex 2a is probably highly unstable and instantaneously
disproportionated into a free pma ligand and dinuclear complex 2b. Multinuclear NMR spectroscopy
and X-ray single crystal structure determination revealed that 2b adopts a structure with two Rh(I)
metal centers hosted by a dianionic (pma–2H)2– ligand, and with the ligand coordinating to Rh2 as an
“aza-allyl” fragment (Figure 1). In contrast to previously reported examples, two-electron transfer
from the (pma–H)– ligand to rhodium(I) along with a second deprotonation step was not observed for
this system. Instead, substantial delocalization of the “excess” electrons over the “aza-allyl” fragment
and the pyridine ring occurs. The inability of this system to form a rhodium(-I) complex is proposed to
result from the lack of an extra coordinating site on the ligand to stabilize the reduced metal in a
149
Summary
preferred tetrahedral coordination geometry. This study, in combination with previous results from
our group, provides useful details about the electron transfer chemistry of these pycolyl-amine-type
ligands, not only showing the reducing ability of NH-deprotonted pma ligands but also their versatile
nature and the subtle dependency of ligand-to-metal electron transfer on a number of factors.
Figure 1. Transformations of 1a and 1b upon NH deprotonation.
In Chapter 3, catalytic activity of the iridium complex 1(synthesis described in Chapter 2) in a series
of transfer hydrogenation (TH) related reactions was studied, which showed high activity for especially
α-alkylation of ketones with alcohols via a hydrogen auto-transfer process (Scheme 1), while the
rhodium complex 2 showed no activity. A mechanism involving the Me-pma ligand assisting in hydride
transfer between substrates via pseudo-Meerwein–Ponndorf–Verley (MPV) type transition states was
proposed (Figure 2) and examined with experimental model studies and DFT calculations. Labelling
experiments indicated that hydride transfer from the substrate (isopropanol) to the corresponding
imine ligand can readily occur under mild conditions, and the reverse process was indirectly proven by
substrate (acetophenone and trans-chalcone) reduction by complex 1a in the presence of a protic
reagent. DFT calculations reveal that the TH reaction pathway can proceed via “reversible hydridestorage” at the imine ligand, and the corresponding substrate-to-ligand and ligand-to-substrate
hydride transfer steps both proceed with moderate barriers (Figure 2). This new ligand cooperative
pathway can be competitive with other mechanisms proposed in earlier studies (e.g. traditional innersphere TH mechanisms).
Scheme 1. α-alkylation of ketones with alcohols catalyzed by 1.
Figure 2. Proposed pseudo-MPV transition
state.
150
Summary
In Chapter 4, the preparation of a new PNN ligand is described, which is derived from a unique
combination of the well-known “proton responsive” 2-(phosphinomethyl)pyridine structure with the
pma type ligand fragment studied in Chapter 2 and 3. This PNN pincer ligand was further used to
prepare the corresponding rhodium carbonyl complex 3. Complex 3 is able to undergo sequential
deprotonation and “hydride-loss” (or proton loss combined with two-electron oxidation) to form
complex 4, which bears a very interesting ligand structure incorporating both a nucleophilic phosphine
arm and an electrophilic imine arm (Figure 3). Complex 4 shows interesting “Janus-type” ligand
reactivity in the splitting of sulfonamide substrate oTsNH2, leading to formation of a (fully
characterized) rare hemi-aminal complex 5, where the proton and the sulfonamido group are each
stored at one of the two arms of ligand. Notably, the sulfonamide N atom replaced the aniline atom as
the coordinating donor bound to rhodium (Scheme 2). Further reaction of complex 5 with
iodomethane (believed to occur on the metal center via oxidative addition) leads to formation of the
methylated sulfonamine oTsNHMe and imine complex 7. Subsequent deprotonation of complex 7
regenerates compex 4.
Figure 3. Synthesis of complex 4 from complex 3 via sequential deprotonation and “hydride-loss”. The figure on
the right shows the HOMO (solid surface) and LUMO (wireframe) of complex 4.
Scheme 2. Hypothetical catalytic cycle composed of stoichiometric steps (one turnover) involving monomethylation of o-TsNH2 with MeI mediated by complex 4.
151
Summary
In Chapter 5, the activity of a series of PNN rhodium carbonyl complexes (based on similar PNN
scaffolds as those described in Chapter 4) in catalytic synthesis of ketenes and ketene imines from
carbene precursors is described. The purpose of using rhodium-based catalysts for these reactions is
to develop a catalyst that can both catalyze ketene formation from carbene precursors and mediate
the following coupling reactions, as this might eventually allow the development of enantioselective
protocols in follow-up studies. Rhodium is expected to have a higher affinity to bind ketenes
compared to the reported catalysts for this reaction, and is thus expected to give rise to inner-sphere
coupling reactions between the in situ formed ketene and the nucleophile/imine. In the catalytic
reactions described in Chapter 5, ketenes were generated in situ by carbonylation of a diazo
compound or a sodium N-tosyl hydrazone salt as the carbene precursor, mediated by the PNN
rhodium carbonyl complexes 8 and 9 under basic conditions (Scheme 3). Trapping of the thus formed
ketenes with an amine or imine in one-pot reactions leds to formation of amides and β-lactams,
respectively, in moderate to high yields. A similar reaction using an isocyanide instead of carbon
monoxide led to formation of a ketene imine, providing the first catalytic example for the synthesis of
ketene imines from a carbene precursors (Scheme 4). The ketene formation and the subsequent [2+2]
Staudinger reaction steps were evaluated with DFT, showing that diazo activation to form the
metallocarbene species is the rate-determining-step, with the neutral amido form of the complex
being more active than the cationic amine form, in agreement with experimental observations. Outersphere CO insertion is computed to be slightly favored over the inner-sphere mechanism, but followup coupling mechanisms are very similar in energy for both pathways.
Scheme 3. Amide and β-lactam synthesis from in situ generated ketenes generated by Rh-mediated
carbonylation of carbene precursors using the PNN rhodium carbonyl complexes 8 and 9.
Scheme 4. Catalytic ketene imine synthesis by coupling of a carbene precursor and an isocyanide
catalyzed by PNN rhodium carbonyl complex 8.
In Chapter 6, catalytic studies of a series of PNN rhodium carbonyl complexes in the
hydroformylation of styrene and 1-octene is presented. Complexes 8 and 9 (see also Chapter 5) show
152
Summary
good activity and regioselectivity (branched over linear; b/l) in the hydroformylation of styrene. The
presence of a base is critical for achieving good catalytic performance. Very high selectivity of the
branched aldehyde product (b/l ratio > 99) was obtained with 1 eq. of KOtBu as the base. When K2CO3
was used as the base, the aldehyde yields were somewhat higher, while the selectivity is still high but
lower than obtained with KOtBu as the base. A catalytically active intermediate 10 was
spectroscopically observed using 8 as the precatalyst, which is even able to catalyze the reaction at
room temperature with high regioselectivity, showing the unusual coordination chemistry of the PNN
ligand scaffold. High pressure NMR spectroscopy suggested that species 10 exists in a monomeric RhI
form hosted by a pyridine-dearomatized PNN ligand or in a dimeric Rh0 form with a P,N-bidentate
ligand. High pressure FT-IR spectroscopy and DFT calculations suggested that species 10 is a
mononuclear [RhI(CO)3(2-P,N-PNN*)-1] species, probably adopting a trigonal bipyramidal coordination
geometry with the phosphorus donor and the pyridine nitrogen donor occupying an axial and an
equatorial position, respectively. A mechanism was proposed based on the structure of species 10, in
which the cooperative ligand assists in the H2 activation step (Scheme 5).
Scheme 5. Proposed mechanism for the PNN-Rh catalyzed hydroformylation of styrene, featuring
ligand cooperativity in the H2 splitting step.
Overall this Thesis shows that cooperative ligands provide versatile reactivity platforms, thus
stimulating follow-up research. The unusual ligand assisted hydride transfer mechanism proposed in
Chapter 3 for transfer hydrogenation and hydrogen auto-transfer-type substrate activation processes
mediated by pma Rh/Ir complexes is noteworthy in this perspective, and application of related
complexes with similar ligand structures might well lead to unexplored new reactivity patterns in
organic synthesis. Also the unusual “Janus-type” PNN-ligands described in Chapter 4 provide
interesting possibilities. The combination of an electrophilic and a nucleophilic ligand arm in one
cooperative ligand was shown to provide a versatile ligand-based reactivity platform for the activation
of a sulfonamide substrate, and the cooperative ligand allowed reversible “storage” of the activated
substrate in the ligand backbone to enable a subsequent metal-based substrate activation proces. A
pseudo-catalytic reaction was achieved based on integration of the multiple processes involved,
thereby demonstrating that similar dual-mode ligand reactivity patterns in general may be a viable
153
Summary
approach in organometallic catalysis. The catalytic activity of similar Rh-PNN complexes in ketene and
ketene imine synthesis from carbenes, as well as in styrene hydroformylation, further underlines the
versatility of the PNN-Rh platform in homogeneous catalysis. We therefore believe that further
mechanistic studies and ligand modifications will lead to further improvements of catalytic efficiencies
and development of new catalytic reactions.
154